[PATCH 2/2] gdi32: Clip font glyphs to fit within text metrics.

[Ken Thomases]

On Apr 22, 2013, at 6:40 PM, Sam Edwards wrote:

> On 04/19/2013 10:32 AM, Max TenEyck Woodbury wrote:
>> As I understand it, some fonts deliberately have glyphs larger than
>> their metrics bounding boxes.  Clipping them is almost certainly not a
>> good idea.
> 
> Forgive my disbelief, but can you provide an example? It seems like Windows has the same clipping behavior (see my test http://source.winehq.org/patches/data/95792).
> 
> From my understanding, the intent of the ascent metric is that it indicates the maximum ascender on any glyph in the font (and likewise for descent), so the only real reason for the ascent/descent metrics to be wrong is if the font designer made a mistake. (And some tools, like FontForge, will automatically set the ascent/descent metrics correctly for you on export.)
> 
> I can't think of any reason why a font author would want to create a font with an invalid ascent/descent metric.

I can't speak to how Windows would render it, but one such font is Zapfino.  It's an "exuberant" calligraphic font with lots of flourishes, some of which have strokes extending into the line above or below.
http://en.wikipedia.org/wiki/Zapfino
https://www.google.com/images?q=zapfino

-Ken